Normalize audited point semantic mappings

[jscott3201]

Summary

• align AHU, VAV, and SYS ontology version echoes with the immutable pin record
• correct the audited VAV/SYS observable-versus-actuatable, quantity, unit, medium, and aspect metadata
• add deterministic network-free point semantic lint, adversarial tests, and CI enforcement
• preserve all 14 dictionary paths, 183 canonical point keys, point ordering, and runtime artifacts

Validation

• python3 -m unittest tools.lint.tests.test_point_semantics -v (16 tests)
• python3 tools/lint/point_semantics.py (14 dictionaries, 12 bounded reviewed mappings)
• python -m unittest discover -s tools/lint/tests -v (86 tests)
• routine schema, semantic, and catalog lints
• pinned G36 source inventory check (431 release, 693 development files)
• cargo test --manifest-path tools/verify/Cargo.toml (6 tests)
• routine and full fault verification through the pinned engine
• dataset harness (32 tests) and tiny FPU replay
• book generation and mdBook build

Boundaries

This does not rename or relocate points, create the shared zone dictionary, normalize role or derived-aggregate aspects, add production derivation manifests or routine profiles, certify arbitrary external ontology terms, validate building instances, or modify Engine or Studio.

[jscott3201]

Cycle 1 review target: e1a6f66d2228635119e5d5cdf67387ba197b395e.

PASS — both independent reviewers are blocker/major-clean. The audited mappings, pin-derived namespace checks, deterministic JSON failure behavior, direction rules, bounded mapping guard, CI wiring, and downstream point-key flows were reviewed at this exact head.

Exact-head build and verify checks completed successfully. All 14 dictionary paths and 183 canonical point names remain unchanged. No minor or nit findings were carried forward.